Childhood obesity in the United States has dramatically increased in the past decade. The proportion of children exceeding the 95th and 85th percentiles for body mass index (BMI) is among the greatest in Mexican- Americans. Despite the high prevalence of obesity among Hispanic children, the genes underlying the heightened susceptibility to childhood- onset obesity in Hispanic populations have not been investigated. The specific goal of this project is to test the hypothesis that a number of genes, each with a measurable effect on the expression of childhood obesity, can be identified by the use of a systematic genomic screen. The specific aims are: l) to identify and phenotype 300 obese Hispanic probands (ages 4-18 y) and their biological parents and siblings, 2) to construct a 10 cM map for 1600 Hispanic individuals to be used in a genome-wide scan for loci that affect quantitative phenotypes of adiposity and energy expenditure using high-throughput genotyping techniques, 3) to perform a multipoint genome scan to find and localize QTLs that influence quantitative variation in adiposity. and energy expenditure in children by performing variance component linkage analysis, and 4) to use multivariate quantitative trait linkage analysis to test whether QTLs localized in Aim 3 have measurable pleiotropic effects across phenotypes. Our target sample will include 1600 genotyped individuals dispersed over 300 nuclear families with a minimum of three children, ascertained on the obese proband using a bivariate scheme (i.e., >95th percentile for BMI and >85th percentile for fat mass). Phenotyping will include anthropometry and body composition, as well as factors associated with the development of obesity: energy partitioning during growth, energy expenditure, physical fitness and activity. hormones, metabolites, and neurotransmitters. Anthropometry and body composition measurements will be repeated after 1y to determine body weight and fat change in the children. Approximately 360 hyper-variable STR markers will be typed for each individual to produce a 10 cM genome map. Multipoint linkage analysis using variance components methods will be applied to nuclear family data to search for QTLs influencing obesity-related phenotypes in Hispanic children. We will test the null hypothesis that the additive genetic variance due to a QTL equals zero (no linkage) by comparing the likelihood of this restricted model with that of a model in which the variance is estimated. This QTL method will be implemented in the program package SOLAR using estimation procedures from FISHER. Lastly, we will test for pleiotropic effects of obesity-related QTLs across phenotypes.